Possible to calibrate a mic using REW and another mic?

[kemmler3D]

I have a UMIK-2 and a Behringer ECM-8000, maybe two if I can find the other one. The Behringer is probably just okay, but more importantly, I have no calibration file or measurements for it. Still, it might be useful to use it for certain types of measurement.

If I take a measurement with the UMIK and the Behringer at the same location, could I calibrate (perhaps a strong word) the Behringer by compensating it to match the UMIK response?

In other words, compare the traces and EQ the behringer to match the UMIK?

Seems intuitive but is there anything to watch out for when doing this, aside from the general sense that it won't be truly reliable?

I'm thinking this might be useful for DIY speaker building or room EQ, I am not going to go out and start an acoustics business on the back of the ECM-8000 or anything.

 

[staticV3]

Professionally, this is done using a setup like this:



If you'd like to "calibrate" your mic using some speakers measured in-room, then I would strongly recommend not doing Mic1 vs Mic2 single sweep at the same location, since even the tiniest difference in position can affect the measured response massively. It's just not a repeatable enough method.

Instead, I would do an MMM with Mic1 vs MMM with Mic2 (PDF Download Link).
Here are the relevant settings you'd use in REW:

Set levels, low cut, and high cut as you see fit. 128k FFT may also be a bit overkill.

Mind you, MMM still leaves some room for run-to-run variation, but it's leaps and bounds more repeatable than a single in-room sweep.

Also of note is that miniDSP's cal files for their mics aren't perfect:

So while you may compensate your Behringer to match your miniDSP, it ends there.

Also also of note is that microphones drift over time. So you may find that as time goes on, the difference in response of the two mics will grow further and further apart.

 

[kemmler3D]

Thanks, that makes sense. I realized that placement would be a big problem, didn't think to do MMM instead. I was imagining using some kind of jig to ensure placement was millimeter-exact. MMM should be easier. I know the accuracy of the ECM will always be some fraction of the accuracy of the UMIK, that's OK since I probably won't be in a situation where I only have the ECM to rely on.

 

[Blumlein 88]

You can do it with sweeps. Put it about 1.5 to 2 meters from your speaker. Be very, very precise with placement of the mike diaphragm. Be wary of low frequency noise you may not notice varying between measures. So do several with each mike until you get some consistency.

 

[kemmler3D]

You can do it with sweeps. Put it about 1.5 to 2 meters from your speaker. Be very, very precise with placement of the mike diaphragm. Be wary of low frequency noise you may not notice varying between measures. So do several with each mike until you get some consistency.

Would it make sense to do the sweeps nearfield, to avoid reflections and other random nonsense?

 

[Blumlein 88]

Would it make sense to do the sweeps nearfield, to avoid reflections and other random nonsense?

Unless using a single speaker the closer you get, the more picky position is because of sound coming from 2 or more drivers. It does help a little to put mic and speaker near middle of the room to get reflections as far away as possible.

 

[kemmler3D]

Unless using a single speaker the closer you get, the more picky position is because of sound coming from 2 or more drivers. It does help a little to put mic and speaker near middle of the room to get reflections as far away as possible.

For the purposes of calibrating, one speaker and a 3D-printed stand-off for the mic is doable.

 

[Blumlein 88]

I suggest doing it with a regular mic stand just for the experience. A 3D mic holder would seem to fix the issue, but microphones are different sizes, fit into the holder a little different and determining with high precision where the diaphragm is means such a mic holder doesn't fix all issues. It likely would help.

Not quite what you are doing, but I've posted results of using published microphone FR graphs as cal files. So here are a couple results.

Is it possible to flatten out a non measurement microphone?

I was wondering if one could flatten out a microphone by inverting the frequency response found in a graph. I would like to get a taste of the result without having to buy an expensive microphone for this... There's a Superlux microphone, which is pretty cheap, but requires an audio interface...

www.audiosciencereview.com

In this case I was using an LSR 305, I've got a pair of fairly large rooms that come together at an interior corner. So the speaker was placed at a 45 degree angle to that corner and walls causing direct reflections were far away in this case. Smoothing was 1/6th octave in these graphs, but they hold up well even at 1/12th.

It would be possible to create a cal file for exact results. I've not tried doing that combined with dismounting and remounting a mic to be calibrated to see how much a difference one might have as a result. The results I've gotten are not perfect, but I'd think pretty usable. You aren't going to get NIST quality results doing it this way.

While I'm thinking of it, mics have different sensitivities. I didn't compensate for those as I kept the signal to the speaker the same. The graphs you see I simply moved the graph in REW to line them up. One could even quibble a bit with my alignment by a db or so.

Also the Umik 2 has a larger diaphragm than the ECM 8000, so its off axis response will differ a bit at the higher frequencies.

 

[SIY]

Be very, very precise with placement of the mike diaphragm.

I use a string and a pointed weight hung from the ceiling to locate the mike, then tie it aside for the measurement. That gets me better than a mm of repeatability. This can be checked by measuring the reference mike again after measuring the DUT, comparing it to the reference measurement, and seeing how close the measurements are. Shoot for 0.05dB or better, which is surprisingly do-able.

 

[Blumlein 88]

I use a string and a pointed weight hung from the ceiling to locate the mike, then tie it aside for the measurement. That gets me better than a mm of repeatability. This can be checked by measuring the reference mike again after measuring the DUT, comparing it to the reference measurement, and seeing how close the measurements are. Shoot for 0.05dB or better, which is surprisingly do-able.

That is a good idea, thanks. I'd used a couple of convenient sticks of PVC. Put them along the based board of the wall corners and created a triangle where the mike stand pole would be located. The mic stand can wobble a bit, so your method would be an improvement I think.

 

[fluid]

If I take a measurement with the UMIK and the Behringer at the same location, could I calibrate (perhaps a strong word) the Behringer by compensating it to match the UMIK response?

There is a good article here that describes that process and the accompanying 90 deg calibration article. It works surprisingly well with a single fullrange driver in a cardboard baffle.

https://johnr.hifizine.com/2012/12/diy-mic-calibration/

REW's trace arithmetic options and outputting the measurement to text could cut down the manual labour a lot.

 

[Blumlein 88]

There is a good article here that describes that process and the accompanying 90 deg calibration article. It works surprisingly well with a single fullrange driver in a cardboard baffle.

https://johnr.hifizine.com/2012/12/diy-mic-calibration/

REW's trace arithmetic options and outputting the measurement to text could cut down the manual labour a lot.

You can do better than that. No need to match the level of the mikes as long as the speaker level is consistent.

Open up the calibrated mic sweep in REW. Then the 2nd mic. Under controls first choose time align. Then choose align SPL, I usually go with 1 khz and select 80 db SPL.

There is an easier way built into REW which I just now realized. Should have been obvious. I detail that easier, and more accurate procedure in this post. People responding below pointed it out. So I was dumb not to know this.

Dayton Audio EMM-6 as good as UMIK-1 for someone who already has an audio interface?

That's not surprising. Even if the housing wasn't so similar, there are not that many mic companies who make their own capsules. That's true for any type of microphones. Two companies I know of that make high-end measurement capsules are Brüel & Kjaer and Microtech Gefell, and those are orders...

www.audiosciencereview.com

Then under Trace arithmetic do A/B. Down near zero will be your calibration curve. Take a screen shot of that.

Then use Web Plot Digitzer. This will let you take that graph and turn it into a .csv file which REW can use for a cal file. You can have several hundred points with more precision and resolution than doing it manually. While being far less tedious. You could even use a graph with 1/12th or 1/24th octave smoothing if you wish.

WebPlotDigitizer - Extract data from plots, images, and maps

automeris.io

Do read and watch the tutorials. Once you have used it once or twice it actually is a piece of cake. Just a bit unusual when you first look at it.

The manual method would work if needed. This is just much more precise.

 

[fluid]

You can do better than that.

It was an example to show the thinking behind it, not a recipe to follow perhaps I wasn't clear. Within the latest REW there are enough tools to output a suitable text file without needing to take a screenshot. But I much prefer VituixCAD's SPL Trace tool if I ever have the need to do that.

 

[OCA]

 

[Weeb Labs]

You can do better than that. No need to match the level of the mikes as long as the speaker level is consistent.

Open up the calibrated mic sweep in REW. Then the 2nd mic. Under controls first choose time align. Then choose align SPL, I usually go with 1 khz and select 80 db SPL.
Then under Trace arithmetic do A/B. Down near zero will be your calibration curve. Take a screen shot of that.

Then use Web Plot Digitzer. This will let you take that graph and turn it into a .csv file which REW can use for a cal file. You can have several hundred points with more precision and resolution than doing it manually. While being far less tedious. You could even use a graph with 1/12th or 1/24th octave smoothing if you wish.

WebPlotDigitizer - Extract data from plots, images, and maps

automeris.io

Do read and watch the tutorials. Once you have used it once or twice it actually is a piece of cake. Just a bit unusual when you first look at it.

The manual method would work if needed. This is just much more precise.

Another way to do this without the need for an external utility is to ensure that the A/B result is the only one enabled and then choose File -> Export -> Export Measurement as Text. This will produce a valid calibration file.

 

[LancerFIN]

Yes it is absolutely possible to make a calibration file for another microphone using calibrated microphone.

This is very good video tutorial on how to make new calibration file for microphone using REW and UMIK-1

A friend of mine bought Omnitronic MIC MM-2USB microphone. Similar to UMIK-1 but costs only 50€. No calibration file included.

https://www.musikhaus-korn.de/en/omnitronic-mic-mm-2usb-kondensator-messmikrofon/pd/138024

I followed the video tutorial above to make him a calibration file for his Omnitronic. After calibration there was virtually no difference between the measurement results of UMIK-1 and Omnitronic MM-2USB.

 

[sam_adams]

Contact a professional.